Groq-Optimized Aider Configuration

A streamlined Aider configuration that leverages Groq's ultra-fast inference infrastructure with intelligent model aliases for efficient AI-assisted coding workflows.

What This Configuration Does

This Aider setup provides:

**Default model**: Groq's GPT-OSS 120B model for balanced performance

**Quick model switching**: Single-character aliases for instant model changes

**Privacy-first**: Analytics disabled by default

**Flexible linting**: Auto-lint disabled for manual control

**Clean git workflow**: Gitignore modifications disabled

Model Aliases

The configuration includes four convenient aliases:

`i` → Llama 3.1 8B Instant (fastest, lightweight responses)

`g` → GPT-OSS 120B (default, balanced quality/speed)

`f` → GPT-OSS 20B (faster inference, lighter model)

`k` → Kimi K2 Instruct (Moonshot AI's instruction-tuned model)

Setup Instructions

1. **Install Aider** if not already installed:

```bash

pip install aider-chat

```

2. **Create the configuration file** at `~/.aider.conf.yml`:

```yaml

model: groq/openai/gpt-oss-120b

analytics-disable: true

auto-lint: false

gitignore: false

alias:

- "i:groq/llama-3.1-8b-instant"

- "g:groq/openai/gpt-oss-120b"

- "f:groq/openai/gpt-oss-20b"

- "k:groq/moonshotai/kimi-k2-instruct-0905"

```

3. **Set your Groq API key**:

```bash

export GROQ_API_KEY="your-api-key-here"

```

Get your API key from [Groq Console](https://console.groq.com/)

4. **Launch Aider** in your project directory:

```bash

aider

```

Usage Examples

Switch models during a session

Within an Aider session, change models using the aliases:

```

/model i # Switch to Llama 3.1 8B for quick responses

/model g # Switch to GPT-OSS 120B for balanced work

/model f # Switch to GPT-OSS 20B for faster iteration

/model k # Switch to Kimi K2 for instruction-following tasks

```

When to use each model

**Llama 3.1 8B Instant (`i`)**: Simple refactoring, quick questions, code formatting

**GPT-OSS 120B (`g`)**: Complex feature implementation, architectural decisions

**GPT-OSS 20B (`f`)**: Iterative debugging, rapid prototyping

**Kimi K2 (`k`)**: Following detailed specifications, structured outputs

Configuration Options Explained

`model`: Default model used when starting Aider

`analytics-disable: true`: Prevents usage data collection

`auto-lint: false`: Manual control over when linting runs (prevents auto-formatting conflicts)

`gitignore: false`: Aider won't modify your .gitignore file

`alias`: Short commands for quick model switching without typing full model names

Important Notes

All models run on Groq's infrastructure for ultra-fast inference (typically <1 second for first token)

Groq API key required (free tier available at console.groq.com)

Model availability depends on Groq's current offerings—verify models exist before use

The configuration assumes Groq as the default provider; adjust model strings if using different providers

Analytics are disabled for privacy; remove this line if you want to support Aider development metrics

Customization

To add your own model aliases, extend the `alias` section:

```yaml

alias:

- "short-name:provider/model-identifier"

```

Example for adding Claude via Anthropic:

```yaml

alias:

- "c:anthropic/claude-3-5-sonnet-20241022"

```

Then use `/model c` within Aider to switch to Claude.

Groq-Optimized Aider Configuration

Groq-Optimized Aider Configuration

What This Configuration Does

Model Aliases

Setup Instructions

Usage Examples

Switch models during a session

When to use each model

Configuration Options Explained

Important Notes

Customization

Reviews (0)